The Kenny et al article entitled "Electron Tunnel Sensor Technology", presented at the first national conference and exhibition of NASA's technology for transfer in November of 1990, describes a micromachined servo accelerometer that utilizes a tunnel current sensor. The accelerometer is micromachined from silicon and includes a cantilever spring with an integral tip. A gold film is deposited over the tip to form a tunnel current electrode. A gold film is also deposited over the cantilever spring to form an electrostatic drive electrode. The inner rectangular area of the folded cantilever spring, here referred to as a proof mass, can be deflected relative to the outer segments, here referred to as a frame, by application of an electric potential between the drive electrode and a corresponding drive electrode disposed on another component of the accelerometer.
Once assembled, a bias voltage is applied to the electrostatic drive electrodes to close the electrodes and drive the proof mass to a servo null position at which a tunnel current having a predetermined value is established. Active regulation of the tip-electrode separation is carried out using feedback control.
Operation of the device as an accelerometer may be achieved in either of two ways. In the first approach, denoted as open loop, acceleration is measured at frequencies above the feedback loop bandwidth in accordance with a predetermined mathematical relationship. In the second approach, denoted as closed loop, acceleration is measured for all frequencies less than the feedback loop bandwidth. In this case, an acceleration displaces the proof mass. The displacement results in a corresponding change in the tunnel current from its predetermined value. A feedback loop responds to the change in the tunnel current by adjusting the voltage potential between the drive electrodes so as to return the proof mass to its servo null position. The variation in the voltage from its bias value is used to calculate the acceleration value since the acceleration value is a function of the voltage variation.
The drive electrodes of the accelerometer can only apply an attractive force which draws the electrodes toward one another. As a result the electrostatic drive can provide the required servo rebalance force only when the acceleration is in a direction which drives the electrodes apart from one another. When an acceleration is applied in the opposite direction in which the electrodes are driven toward one another, the voltage difference between the electrodes is decreased thereby decreasing the drive force. The flexures which connect the proof mass to the frame then provide an elastic force to return the proof mass to its servo null position. Without acceleration, the elastic force provided by the flexures must at least be equal to the rebalance force required to reposition the proof mass to its servo null position upon application of full scale acceleration. Likewise, the electrostatic drive must be capable of providing enough force to drive the proof mass to its servo null position upon application of full scale acceleration. To provide the necessary dynamic response, the forces which the electrostatic drive and the flexures are respectively capable of providing must exceed the minimum force required to reposition the proof mass to its servo null position upon application of full scale acceleration.
The accelerometer of the Kenny et al article is deficient in several respects. Both the stiffness of the flexures and the position at which the flexures hold the proof mass when no forces are applied (i.e., the mechanical null position) can change with aging, temperature, and other environmental effects. Such changes produce a corresponding change in the drive force required to keep the proof mass at its servo null position and thus effect the bias voltage applied to the drive electrodes. Also, changes in the values of the components of the electrical circuitry of the feedback loop can cause a change in the servo null position to a position where the elastic force from the flexures is different from the elastic forces of the original servo null position. This alters the drive force required to hold the proof mass so that the drive force at the new null position is different from the drive force required at the original null position. Since the drive voltage is used to calculate the acceleration value, such changes cause a change in the acceleration signal bias.
A further deficiency of the Kenny et al accelerometer is that the force on the electrostatic drive electrodes is approximately proportional to the square of the charge on the drive electrodes. There is thus a non-linear relationship between the acceleration value and the drive voltage variations used to measure the acceleration.